<lgierth>
yes yes -- but i think david introduced some custom adhoc sharding to registry-mirror to circumvent this 1MB object size limit
<lgierth>
Kubuxu: would these malformed block names potentially crash gc?
Boomerang has quit [Quit: Leaving]
<kpcyrd>
jbenet: is it possible to `ls` those?
<jbenet>
wa7son: are you guys pushing the webrebels videos to youtube? would like to watch them at 2x
<jbenet>
(also linking to specific time code...)
<jbenet>
kpcyrd: yeah of course :)
ipfs-newbie has joined #ipfs
<ipfs-newbie>
ne 1 here?
<lgierth>
ipfs-newbie: yes, exactly 582 people -- you can just ask, without asking to ask :)
vakla has joined #ipfs
<ipfs-newbie>
I usually wait when I pop in a channel, but didn't see any discussion in 30 seconds. With this many people here, I figured it was a lot of afk, bots, and stuff
<kpcyrd>
jbenet: how close are those to being available in go-ipfs?
ashark has joined #ipfs
<lgierth>
good question! sharding spec jbenet? :P
<ipfs-newbie>
my question, is I get the basic premise of ipfs. is there any benefit or reasons to run an ipfs node, other than you want to support the network?
<lgierth>
you might want to *use* ipfs? :)
<voxelot>
im always lurking, no life, what's up ipfs-newbie
<kpcyrd>
btw, is a node that does not store objects useful to the ipfs network?
<ipfs-newbie>
ok, I think I'm a little tained because I just read a post on steemit from an application level how they say they will use ipfs on their backend. someone said there may not be enough ipfs nodes yet for that to be 100% effective
<lgierth>
kpcyrd: yes, for IPNS, and peer routing
<jbenet>
kpcyrd lgierth: i've passed it onto whyrusleeping. we agreed on using HAMTs, which are great persistent data structs.
<jbenet>
"there may not be enough nodes..." i dont think they understand how IPFS works.
<lgierth>
jbenet: a great! that's cool -- sharding's coming!
<lgierth>
ipfs-newbie: part of the idea of ipfs is that there will be no backend or server anymore -- right now we're working on ipfs in javascript which can run in browser and websites, and it's already partly usable
<ipfs-newbie>
I'll give you a scenario similar you want to download say an old tv show off of bittorrent, only 10 seeders. meanwhile something else has 1800 seeders
stoogenmeyer has joined #ipfs
<ipfs-newbie>
in ipfs could a similar situation occur? only 10 seeders of a particular jpg
<jbenet>
<3
<kpcyrd>
jbenet: I would add a workaround with `ipfs object patch add-link` and hope I stay below 1MB until it's ready
<lgierth>
ipfs-newbie: we usually don't do copyright infringement here -- in general your ipfs node only stores what you add yourself, or request yourself. there will be tools for multiple nodes to agree on keeping a set of things available (ipfs-cluster)
<kpcyrd>
ipfs-newbie: ideally, if you are interested in content staying around, you can `ipfs pin` it on some of your nodes
<lgierth>
and there's also going to be filecoin, a cryptocurrency for rewarding nodes for keeping things available
<ipfs-newbie>
i know about the copyright thing. i only used bittorrent because its p2p and most everyone knows about how seeders work on that
<voxelot>
file availability is a frequent question here, excited to work on filecoin.
<ipfs-newbie>
AHH you just answered my question jbenet
<ipfs-newbie>
there is a reward for participating. many people will want to seed, store, share and forward as much as they can. :)
<jbenet>
ipfs-newbie: :) search the faq -- usually someone has already asked the common questions.
<jbenet>
ipfs-newbie: yes and no, the reward is a _different_ protocol.
<jbenet>
(layered on top)
<ipfs-newbie>
thanks. reading faq now. git links always confuse me into readme files that are uncomfortable on the eyes. I normally stray from them. human habit.
<kpcyrd>
ipfs-newbie: I'm working on exporting my mirror server into ipfs, so you can access it with https://gateway.ipfs.io/ipns/mirror.rxv.cc/. Should be at least as reliable as using it with regular http. If nobody is seeding the files, the mirror is
<child>
what's ipfs-cluster going to be written in and why isn't it erlang?
<jbenet>
child: pull requests accepted :)
<kpcyrd>
lgierth: I can't, assuming the user doesn't have ipfs installed :)
<lgierth>
that was "gateway." -- i meant you can just go ipfs.io/ipfs/<hash>
<kpcyrd>
ah. `gateway.` is deprecated?
<lgierth>
there are no plans to drop it, but it's only there for historical reasons -- ipfs.io used to be a github page
<jbenet>
child: but more seriously, that's a good question. because core team is not yet erlang pro, and we find Go super maintainable and useful for distributed systems. dont get me wrong, LOVE erlang's model, but the life in Go is quite good.
<ipfs-newbie>
these feels like a version of fideonet in simplistic terms but on steriods and a much larger design. one bbs creates a message board, and its up to all the other fideonet bbs if they want to carry that particular topic / message board too
gsf has quit [Ping timeout: 272 seconds]
gsf has joined #ipfs
<ipfs-newbie>
<-- showing my age
<jbenet>
kpcyrd: we'll lug around gateway.ipfs.io because there are lots of links out there pointing to it. of all people on the internet, We are one of the few people not allowed to break links
<lgierth>
ipfs-newbie: yeah there are plenty of distributed systems from the last 40 years to draw ideas from!
Senji has quit [Ping timeout: 276 seconds]
<lgierth>
jbenet: kpcyrd: aand it's also useful to keep around so people can point CNAME records at it (ipfs.io itself has more than just A/AAAA records)
<ipfs-newbie>
so I am reading about blockstack and namecoin for decentralized dns in conjunction with ipfs. any other leading technologies? decentralized dns is something i am very interested to use with ipfs for true decentralization
<ipfs-newbie>
i just wanted to post this link, so you can see where the ipfs thing got me going and I cant stop until I learn as much as I can. now that you've got me a foundation to start my research, I'll stop asking questions here and concentrate on more educated questions later.
<ipfsbot>
[go-ipfs] whyrusleeping created feat/auto-migrate (+1 new commit): https://git.io/vKfk1
<ipfsbot>
go-ipfs/feat/auto-migrate 3ca23aa Jeromy: Automatically download and run migrations if needed...
<ipfsbot>
[go-ipfs] whyrusleeping opened pull request #2939: Automatically download and run migrations if needed (master...feat/auto-migrate) https://git.io/vKfkH
ylp has quit [Ping timeout: 264 seconds]
ruby32 has quit [Read error: Connection reset by peer]
<kpcyrd>
it looks like ipns is a bit flappy, but after you got the objects locally, it's faster than internet
<whyrusleeping>
kpcyrd: wooo!
ylp has quit [Ping timeout: 264 seconds]
dmr has quit [Ping timeout: 250 seconds]
ipfs-newbie has quit []
<mythmon>
kpcyrd: nice!
Combined2857 has quit [Quit: Leaving]
ylp has joined #ipfs
<kpcyrd>
adding `Server = http://127.0.0.1:8080/ipns/mirror.rxv.cc/archlinux/$repo/os/$arch` to the beginning of /etc/pacman.d/mirrorlist would try to get the file from local ipfs, and if that doesn't work, fall back to the regular mirrors
jo_mo has quit [Quit: jo_mo]
stoogenmeyer has quit [Remote host closed the connection]
stoogenmeyer has joined #ipfs
stoogenmeyer has quit [Remote host closed the connection]
<lgierth>
bittorrent clients can add ipfs support -- interoparibility between the bittorrent mainline dht and ipfs's bitswap is very very unlikely
<lgierth>
whyrusleeping: load average is more interesting that uptime anyhow :)
<lgierth>
props to digitalocean's datacenter for 670 days of uninterrupted power supply :D
pguth has quit [Remote host closed the connection]
pguth has joined #ipfs
<ipfsbot>
[js-ipfs] diasdavid created greenkeeper-rimraf-2.5.3 (+1 new commit): https://git.io/vKf1y
<ipfsbot>
js-ipfs/greenkeeper-rimraf-2.5.3 13cac3e greenkeeperio-bot: chore(package): update rimraf to version 2.5.3...
matoro has joined #ipfs
corvinux has quit [Ping timeout: 276 seconds]
stoogenmeyer has joined #ipfs
<davidak>
whyrusleeping: as i always say, a high uptime is not good. it is bad because it means you are running an old kernel with unfixed vulnerabilitys
<davidak>
(my company has also servers with uptime > 600)
<kpcyrd>
I'm running a modified version of go-ipfs that allows me to put large objects
<davidak>
Kubuxu: that is very high if you have no 240 cpu cores
<Kubuxu>
davidak: no just 24 cores, but it is because we use io throttling
<davidak>
ok
jo_mo has joined #ipfs
herzmeister has quit [Quit: Leaving]
herzmeister has joined #ipfs
reit has joined #ipfs
<davidak>
i like the comparison of pinning with bookmarking, but you also save the content. it would be useful to have a list with page title or file name to find content again
pguth has quit [Remote host closed the connection]
pguth has joined #ipfs
zorglub27 has quit [Quit: zorglub27]
stoogenmeyer has joined #ipfs
f[x] has quit [Ping timeout: 276 seconds]
igi-111 has quit [Quit: Leaving]
ashark has quit [Ping timeout: 276 seconds]
toxip has quit [Ping timeout: 246 seconds]
toxip has joined #ipfs
reit has quit [Quit: Leaving]
pguth has quit [Remote host closed the connection]
pguth has joined #ipfs
dignifiedquire has quit [Quit: Connection closed for inactivity]
Senj has quit [Ping timeout: 252 seconds]
Encrypt has quit [Quit: Quitte]
cemerick has joined #ipfs
r04r is now known as zz_r04r
computerfreak has quit [Remote host closed the connection]
rgrinberg has quit [Ping timeout: 276 seconds]
tmg has quit [Ping timeout: 244 seconds]
tmg has joined #ipfs
rgrinberg has joined #ipfs
Arakela0_ has quit [Remote host closed the connection]